JP2009219064A - Image processing method, image processor, printer, and image processing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus that generates a color conversion table fast. <P>SOLUTION: An LUT for color conversion from device colors of a display 20 to device colors of a printer 40 is generated, a color gamut of the printer 40 is generated with a polygon, and it is determined whether a half-line extending from a gray axis of the color gamut to coordinates after conversion by a color table generating means comes in contact with the polygon plane; and an intersection with the polygon plane determined to be in contact is calculated, the distance from a reference point to the intersection is compared with the distance from the reference point to the coordinates after the mapping to determine whether the coordinates after the conversion by the color conversion generating means are in the color gamut, and coordinates determined to be outside the color gamut are re-mapped in the color gamut. The contact decision making targets the polygon plane decided to include an intersection with the coordinates obtained by last color conversion and a polygon plane enclosing a circumference thereof. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an image processing method, an image processing apparatus, a printer, and an image processing program, and in particular, sets a correspondence relationship between a first color space of a first device-dependent color and a second color space of a second device-dependent color. And an image processing method, an image processing apparatus, a printer, and an image processing program for converting a first color in the first color space into a second color in the second color space based on the correspondence.

Generally, color gamuts (color reproduction gamuts) differ among image input devices such as digital still cameras and scanners, output devices such as printers, display devices such as displays, and the like. Some image data have different color gamuts. When the color gamuts are different, when image data expressed in one color gamut is displayed in the other color gamut, color matching processing (color gamut conversion processing) is performed to maintain gradation and color reproducibility. Done.
JP 2007-189632 A

  The color matching process is generally performed using a color conversion table (LUT) that associates representative grid points in each color gamut. Such an LUT is performed by associating both profiles in a predetermined connection space while referring to the ICC profile of the input device and the ICC profile of the output device, as described in Patent Document 1, for example. However, such processing takes time, and it is desired to increase the processing speed.

  SUMMARY An advantage of some aspects of the invention is that it provides an image processing method, an image processing apparatus, a printer, and an image processing program capable of speeding up the creation of a color conversion table.

In order to solve the above problems, in the image processing method of the present invention, a correspondence relationship between the first color space of the first device-dependent color and the second color space of the second device-dependent color is set, and the correspondence relationship is set. An image processing method for converting a first color in the first color space into a second color in the second color space based on a color conversion table setting step, an outer shell setting step, and a detection step And an inside / outside determination step and a mapping step. In this configuration, in the color conversion table setting step, a color conversion table to be referred to when converting the first color to the second color is set based on a predetermined conversion rule. In the outer shell setting step, the outer shell of the second color space is set by combining a plurality of polygonal planes. In the detecting step, a polygonal plane having a half line that intersects with a half line from a predetermined point in the second color space to a color coordinate of the color converted based on the color conversion table is detected. In the inside / outside determination step, the intersection and the color coordinate are compared to determine whether the coordinate is within the outer shell. In the mapping step, when it is determined that the coordinates are not in the outer shell in the inner / outer determination step, the color is associated with the outer shell. In the above configuration, in the detection step, first, a polygon plane detected immediately before and a polygon plane around the polygon plane are detected, and all the polygon planes have no intersection when these polygon planes have no intersection. Is the detection target.
That is, when performing a hit determination between any of the plurality of polygon planes and the half line to detect a polygon plane having an intersection with the half line, not all polygon planes are to be determined. A polygon plane that is likely to be hit is predicted, and the hit determination is preferentially performed on the predicted polygon plane. Thereby, a polygonal plane having an intersection can be detected in a shorter time, and the processing can be speeded up. In other words, the coordinates converted immediately before based on the color conversion table and the coordinates converted immediately after that are highly likely to be color-converted to relatively close coordinate values. On the other hand, a polygonal plane having an intersection can be detected in a short time by setting a polygonal plane adjacent to the polygonal plane centering on the polygonal plane determined to have an intersection. In addition, if the hit determination is not performed on the directly adjacent polygon plane, a method that gradually expands the search area, such as a polygon plane further adjacent to the adjacent polygon plane, is adopted. Good.

  Here, various color spaces can be adopted as the device independent color space, and various absolute color spaces such as Lab, Jab, and XYZ can be adopted. The first and second device-dependent colors (image devices having device-dependent colors) are not particularly limited, and may be any device-dependent color of a device that handles a predetermined image using image data. For example, a display or a printer Device-dependent colors of various image devices such as scanners and digital cameras can be adopted. Of course, the devices having the first and second device-dependent colors may not be separate. For example, in a fax machine, a scanner as a first image device and a printer as a second image device are integrated. However, the present invention can be applied to such a fax machine.

  Further, as a selective aspect of the present invention, the method further comprises a rearrangement step of rearranging the plurality of polygonal planes by brightness and hue angle, and in the intersection detection step, the brightness and hue angle of the color coordinates It is good also as a structure made into a detection target in an order from the polygonal plane approximated to. In other words, if all the polygonal planes that have intersections are not found as a result of hitting the polygonal plane that has been determined to have an intersection with the color-transformed coordinates immediately before, all polygonal planes are determined. At this time, the determination order can be determined based on the hue angle and lightness. Therefore, even if hit determination is executed for all polygon planes, polygon planes having intersections can be efficiently found.

  Further, as a selective aspect of the present invention, in the mapping step, the nearest neighbor point is searched from the polygon plane detected by the detection means and the polygon plane surrounding the periphery, and the nearest store is searched. It is good also as a structure matched. The nearest neighbor point is not necessarily a polygonal plane having an intersection point, but the coordinate value that needs to be mapped in the mapping step is not likely to be far away from the color gamut outer shell, and it is unlikely to be high in the vicinity of the polygonal plane having the intersection point. There is a nearest point with probability. Therefore, this selective aspect is suitable as a technique for determining a mapping destination without unnecessarily increasing the amount of calculation.

  Further, as a selective aspect of the present invention, the predetermined point may be a point on the gray axis of the second color space. Since the gray axis is an axis passing through the approximate center of the color gamut, the half line extending from the gray axis intersects the color gamut shell at a single point with a high probability. Therefore, the possibility of erroneous determination or incorrect coordinates being calculated in the hit determination or intersection calculation is reduced, which is preferable.

  The image processing method described above may be implemented as part of another method, or may be realized as an image processing device provided with means corresponding to each process, or an image processing device provided with a color conversion table created by the method, etc. Including various embodiments. The present invention can also be realized as an image processing system including the image processing apparatus, a program for causing a computer to implement functions corresponding to the above-described method configuration, a computer-readable recording medium storing the program, and the like. . The inventions of the image processing system, the image processing apparatus, the image processing program, and the medium on which the program is recorded also have the above-described operations and effects. Of course, the configurations described in claims 2 to 4 are also applicable to the system, the apparatus, the program, and the recording medium.

  Further, as an optional aspect when the present invention is implemented as an image processing device, the image processing device further includes a graphic board equipped with a GPU (Graphics Processing Unit), and the detection means includes the GPU. The polygon plane having the intersection with the half line is detected, and the inside / outside determination means causes the GPU to calculate a distance from the predetermined point to the color coordinate and a distance from the predetermined point to the intersection. It is good. In other words, by using a GPU that is faster than a normal CPU and optimized for image processing, it is possible to perform hit determination and distance calculation in a short time. Further, an API specialized in a function for obtaining an intersection of a polygon and a ray may be called, and the API may use the GPU to execute a hit determination and a distance calculation.

Hereinafter, embodiments of the present invention will be described in the following order.
(1) Configuration of image processing apparatus:
(2) LUT creation processing:
(3) Printing process:
(4-1) Color gamut conversion processing:
(4-2) Modification of color gamut conversion process:
(5) Smoothing processing of peak value fluctuation between hues:
(6) Speeding up the reverse conversion process:
(7) Color gamut inside / outside determination processing:
(8) Summary:

(1) Configuration of image processing apparatus:
FIG. 1 is a block diagram showing a hardware configuration of a computer that realizes an image processing apparatus according to an embodiment of the present invention. In FIG. 1, a computer 100 includes a control unit 10 having a CPU 10a and a RAM 10b, a ROM 11, a hard disk (HDD) 15 as a large capacity recording device, a USB I / F 13 for connecting an external input / output device, and a display 20. A video I / F 14 to be connected and an operation input device I / F 17 for connecting an operation input device are provided, and each component is connected by a bus 16 so as to be able to communicate with each other. Note that an external printer 40 is connected to the USB I / F 13, and a keyboard 30 a and a mouse 30 b are connected to the operation input device I / F 17.

  The CPU 10a reads out program data from the ROM 11 and the HDD 15 and develops the program data on the RAM 10b, thereby executing an operating system (OS) executed on the BIOS or the BIOS and an application P executed on the OS. In the present embodiment, the OS and application P are stored in advance in the HDD 15 as program data, but the recording medium is not limited to this. For example, a flexible disk or a CD-ROM may be used. Programs recorded on these recording media are read into the computer 100 via a flexible disk drive or a CD-ROM drive, and installed in the HDD 15 or directly executed. Then, it is read onto the RAM 10b to control the computer. The recording medium is not limited to this, and a non-volatile memory such as a flash card can be used as a magneto-optical disk or a semiconductor device. In addition, when an external file server is accessed and downloaded via a modem or a communication line, the communication line becomes a transmission medium and the present invention is used.

  FIG. 2 is a block diagram showing a software configuration of the computer 100. In the figure, in the computer 100, a printer driver (PRTDRV) 21, an operation input device driver (DRV) 22, a display driver (DRV) 23, and a color gamut conversion module 24 are incorporated in the OS. The display DRV 23 is a driver that controls display of image data and the like on the display 20, and the operation input device DRV 22 is a driver that receives a code signal from the keyboard and mouse and receives a predetermined input operation. The color gamut conversion module 24 creates a correspondence relationship between the color gamuts (color reproduction gamuts) of different devices and, based on the correspondence relationship, changes the color gamut color of one device to the color gamut color of another device. Convert to Note that the color and gradation reproducibility can be adjusted when performing this conversion (hereinafter referred to as color gamut adjustment). In addition, the APL 25 can be executed on the OS.

  The APL 25 can read the image data 15a recorded on the displayed HDD 15 to the RAM 12 and execute processing such as retouching. The display DRV 23 can display an image on the display 20 based on the image data 15a read to the RAM 12. Display. When the user operates the operation input device 30, the operation content is acquired via the operation input device DRV 22 and the content is interpreted, and the APL 25 executes printing or retouching according to the operation content. Various processes are performed.

  In addition, the user can reproduce and display an image formed in a predetermined device by performing setting processing by operating the operation input device 30 under execution of the APL 25. This reproduction is performed using the color gamut conversion module 24 in the present embodiment. For example, when displaying the image data created in the sRGB color space, when the color gamut of the display is narrower than the sRGB color space, the portion that protrudes from the color gamut of the display is a predetermined color within the color gamut of the display 20. It can be adjusted to perform color conversion as expressed. Then, the gradation and color expressed in the image data are maintained in the display display. In addition, when color conversion is performed so that the image data is expressed in a predetermined color that can be expressed in the color gamut of the printer 40, a print result in the predetermined printer can be reproduced on the display. Of course, in the reproduction of the print result, the print result may be reproduced after performing color conversion so that the gradation and color of the image data are maintained. Furthermore, the printer 40 can print a color image of the print result displayed with such adjustment. Hereinafter, as an example of color gamut adjustment executed by the color gamut conversion module, an example in which the print result of the printer 40 is adjusted and displayed on the display will be described.

  FIG. 3 shows an example of a screen for performing color gamut adjustment setting processing. In the setting process, device selection on the input side and output side, selection of a profile (Device Model Profile: DMP) corresponding to the device, and a profile (Gamut Map) for specifying a different color gamut adjustment and conversion method (Rendering Intent) Model Profile: GMMP) is selected. When the setting process is completed and conversion of the color gamut is instructed, the color gamut conversion module 24 is activated, and the color gamut of the input side device is converted to the color gamut of the output side device while referring to the set DMP and GMMP. Is created and stored in the HDD 15. The creation of the LUT 15b will be described later. Next, when the user operates the mouse to instruct the display of the print result of the printer 40, an instruction to adjust the color gamut is made, and the image data 15a is transferred to the color gamut conversion module 24. When the image data 15a is transferred to the color gamut conversion module 24, the color conversion process shown in FIG. 3 is started.

  In the present embodiment, the color gamut conversion module 24 creates the LUT 15b and then executes the color gamut conversion process while referring to the LUT 15b. Of course, the color gamut conversion module 24 dynamically creates the LUT. It does not matter. For example, the color adjustment of the color adjustment module 24b may be executed while creating LUTs in the order in which the color coordinates constituting the image data 15a are input. Further, the LUT 15b created in this way is stored in the HDD 15, and when the gamut conversion processing between the same gamuts is instructed, the color conversion processing is performed by referring to the created LUT. Is possible.

Here, the configuration of the color gamut conversion module 24 will be described with reference to FIG. The color gamut conversion module 24 of this embodiment includes the following modules M1 to M4. The XYZ value calculation module M1 converts the color gamut of the input device expressed by RGB values into CIE tristimulus values XYZ values while referring to the DMP of the input device. The XYZ value calculation module M2 converts the color gamut of the output device expressed by RGB values into CIE tristimulus values XYZ values while referring to the DMP of the output device. CA conversion module M3 went forward conversion of the color appearance (CA) conversion formula into XYZ values calculated in XYZ value calculating module M1, M2 calculates the ^J * ^a * ^{b * values,} J ^* a ^{* b *} Determine the color gamut of the input device in the color space. The color gamut mapping module M5 maps the color gamut of the input device determined by the CA conversion module M3 with reference to the GMMP to the color gamut of the output device calculated by the CA conversion module M3. The CA conversion module M4 calculates the XYZ value by performing the reverse transformation of the color appearance (CA) conversion formula on the J ^* a ^* b ^* value calculated by the color gamut mapping module M5, and outputs the color of the output device in the XYZ color space. Determine the area. In the color appearance transformation performed by the respective CA conversion module M3, M4, conversion in consideration of the input or observation conditions specified in the specified DMP for the output device (D _55, D _65, etc. source information) Executed.

  The profiles DMP1 to n and GMMP1 to GMMPn used in the modules M1 to M4 are stored in the HDD 15. Each DMP describes the correspondence for converting the color gamut of the device in the sRGB color space to the colorimetric values and the above observation conditions, and GMMP describes the conversion formulas and conditional expressions for associating the color gamuts. Has been. In this embodiment, the device color reproduction gamut described in the DMP is described as a color reproduction range in the sRGB color space, but may of course be expressed in another color space.

  The color gamut conversion performed using DMP includes forward conversion (corresponding to the AToB1Tag processing of the ICC profile) and backward conversion (corresponding to the BToA1Tagn processing of the ICC profile). The forward conversion is realized by a forward multidimensional LUT or a conversion formula for converting device colors into XYZ values. In reverse conversion, the result of forward conversion is used to reverse the forward multidimensional LUT, optimize the polynomial parameters by regression analysis, etc., and convert the generated XYZ values to device colors. This is realized by a multidimensional LUT or conversion formula to be performed.

In the color appearance conversion, the forward direction conversion of a color perception model such as CIECAM02 is applied to the XYZ values to obtain the color values in the color appearance space. The color value of this color appearance space (for example, J ^* a ^* b ^* value) can be calculated from the CIE tristimulus values XYZ. By using this J ^* a ^* b ^* , it is possible to perform color gamut mapping more suitable for human visual characteristics. The conversion for obtaining the color value J ^* a ^* b ^* of the color appearance space is performed by chromatic adaptation conversion, cone response conversion, and opposite color response conversion, and is converted so as to approach human visual characteristics. In the chromatic adaptation conversion, conversion corresponding to the light source under the observation environment is performed. The color perception model need not be limited to CIECAM02, CIECAM97s, etc., and human color perception parameters J (lightness), C (Chroma), Q (brightness), M (colorfulness), h (huequadrature or hueangle) , H (huequadrature or hueangle) as long as it is a color perception model, other perception models may be used.

(2) LUT creation processing:
FIG. 5 is a flowchart of LUT creation work. Since this operation requires a lot of calculation processing, it is preferable to execute the calculation using a computer.
First, in step S200, the selected DMP1 for the display 20 is acquired and the color gamut of the display 20 is set. For example, if the display 20 is expressed with 256 gradations for each of RGB colors, the XYZ value calculation module M1 converts XYZ values into XYZ values using a known conversion formula while considering all combinations for each color 256 gradations. The color gamut of the display 20 in the space is determined. The CA conversion module M3 further performs color appearance conversion in consideration of the observation conditions described in DMP1 with respect to the XYZ values converted in this way to calculate J ^* a ^* b ^* values. The LUT that defines the correspondence between the RGB values of the display 20 and the J ^* a ^* b ^* values is the first color conversion table. ^Then, J ^* a ^{* b *} gamut mapping module M5 which has received the value, the color gamut by performing the creation of a region encompassing the resulting ^J * ^a * ^{b *} values (for example, a three-dimensional convex hull and polygons, etc.) Set. The color gamut is a three-dimensional solid that encompasses the obtained Jab values.

In step S210, the selected DMP2 is acquired for the printer 40, and the color gamut of the printer 40 is set. If the CMYK gradation data of the printer 40 is expressed with 256 gradations for each color, DMP2 describes which range in the sRGB color space the color gamut that can be expressed with these gradations corresponds to. The XYZ value calculation module M2 determines the color gamut of the printer 40 in the XYZ color space by converting the XYZ values into XYZ values using a known conversion formula while considering the combination of 256 gradations of each RGB color in this range. Further, the CA conversion module M3 performs color appearance conversion in consideration of the observation conditions described in DMP2 with respect to the XYZ values to calculate J ^* a ^* b ^* values. The LUT that defines the correspondence between the RGB values of the printer 40 and the J ^* a ^* b ^* values is the second color conversion table. The ^{J * a} * ^b * gamut mapping module M5 which has received the value, the color gamut by performing the creation of the resulting ^{J * a} * ^b * including the value area (for example, a three-dimensional convex hull and polygons, etc.) Set.

As described above, after determining the color gamuts of the display 20 and the printer 40, color gamut conversion processing is performed in step S220 to map the display color gamut in the printer color gamut in the J ^* a ^* b ^* color space. In the color gamut conversion process in the present application, mapping is performed in which both saturation and brightness are changed at high brightness, and saturation is changed while maintaining brightness at low brightness. Therefore, when reproducing the color of the display in the color gamut of the printer, the gradation of high saturation is improved at high brightness and the color reproduction is improved, and the gradation of the dark portion is improved at low brightness. This color gamut conversion process will be described later.

  Since the correspondence between the display color gamut and the printer color gamut is defined by the above conversion, in step S230, representative points necessary for the interpolation calculation at the time of color conversion are extracted to create the LUT 15b. That is, the RGB values of the display 20 are converted into XYZ values based on the first color conversion table, the color gamut conversion processing is performed on the XYZ values, and the converted XYZ values are converted into the color gamut of the printer 40. By converting the RGB values of the printer 40 by reverse lookup based on the second color conversion table, an LUT 15b that associates the RGB values of the display 20 with the RGB values of the printer 40 in the sRGB color space is created. Since the XYZ values after the color conversion process are not necessarily the points described in the second color conversion table, the XYZ values after the color conversion process are searched from the second color conversion table for a close XYZ value. By specifying the XYZ values at several points surrounding the image and performing interpolation calculation from the RGB values corresponding to the specified XYZ values, the RGB values corresponding to the XYZ values after the color conversion process are determined.

  However, since conversion processing and calculation are performed before the RGB value is determined, the determined RGB value may be out of the printer color gamut in the sRGB color space. Therefore, in step S240, an area including the display 20 color gamut (for example, a three-dimensional convex hull or polygon) is created to set a color gamut in the sRGB color space, and a color gamut inside / outside determination process described later is performed. Become. The printer color gamut in the sRGB color space thus set has a one-to-one correspondence with the display color gamut in the sRGB color space. By converting the image data 15a with reference to the LUT 15b created and set as described above, an image color-converted to the color reproduction range of the printer 40 can be reproduced on the display 20.

  The above-mentioned DMP and GMMP can be added or changed as necessary, and it is easy to change the model of the display 20 or the printer 40, change the ink type or media type used in the printer 40, etc. Can respond.

(3) Printing process:
Here, a printing process using the LUT 15b will be described below. When the PRTDRV 21 receives a print execution instruction, the image data 15a is acquired by the image data acquisition module 21a, and the image data acquisition module 21a activates the color conversion module 21b. The color conversion module 21b is a module that converts RGB gradation values into CMYK gradation values, and converts each dot data of the image data 15a into CMYK dot data. At this time, the color conversion module 21b performs an interpolation operation with reference to the LUT 15b and the LUT 15c stored in the HDD 15. The LUT 15c is an LUT that converts sRGB values into CMYK values. In the present embodiment, the color gamut conversion module 24 creates the LUT 15b in advance, and the created LUT 15b is acquired and used by the color conversion module 21b. The color gamut conversion module 24 may create the LUT 15b and supply it to the color conversion module 21b in response to the request of the module 21b.

  When the color conversion module 21b performs color conversion to generate CMYK gradation data, the CMYK gradation data is transferred to the halftone processing module 21c. The halftone processing module 21c is a module that performs a halftone process for converting the CMYK gradation value of each dot and expressing it with the recording density of the ink droplets, and a head for attaching ink at the converted recording density. Generate drive data. The print data generation module 21d receives such head drive data and rearranges them in the order used by the printer 40. That is, in the printer 40, an ejection nozzle array (not shown) is mounted as an ink ejection device, and in the nozzle array, a plurality of ejection nozzles are arranged in parallel in the sub-scanning direction, and therefore data separated by several dots in the sub-scanning direction. Are used simultaneously.

  Therefore, rasterization is performed in which data arranged in the main scanning direction is rearranged in order so that data to be used at the same time is buffered by the printer 40 at the same time. After the rasterization, print data is generated by adding predetermined information such as the resolution of the image, and is output to the printer 40 via the USB I / F 13. The printer 40 prints the image displayed on the display 20 based on the print data.

(4-1) Color Gamut Conversion Processing Next, the color gamut conversion processing in step S220 will be described in detail. FIG. 6 is a schematic diagram showing a state in which the color gamut of the display and the color gamut of the printer are cut at a predetermined hue angle θ in the J ^* a ^* b ^* space. In the figure, the horizontal axis represents saturation C (= (a ² + b ² ) ^1/2 ), and the vertical axis represents lightness J. In the present embodiment, in the color gamut conversion process, areas outside the printer color gamut and near the outer shell in the display color gamut are mapped near the outer shell of the printer color gamut. It should be noted that the internal area of the display color gamut excluding the vicinity of the outer shell of the printer color gamut is mapped to the same coordinate value.

In the color gamut conversion processing in the present embodiment, the boundary of lightness J ₀ of the coordinate indicating the maximum chroma C _max in the printer gamut cut at each hue angle (C _max, J _0), the high brightness side and the low brightness side Different mapping algorithms are adopted. Hereinafter, (C _max , J ₀ ) is referred to as a peak (cusp), and the brightness of the peak is referred to as a peak value. FIG. 7 is a diagram for explaining a compression algorithm in the present embodiment. The compression algorithm of the present embodiment is an algorithm that compresses both saturation and lightness on the higher lightness side than the peak value while placing importance on maintaining the lightness, such as SGCK (Sigmoidal Gaussian Cusp Knee). Compression algorithm), and on the lower brightness side than the peak value, an algorithm (second compression algorithm) that compresses the saturation while maintaining the brightness, such as BasicPhoto (registered trademark Microsoft), is adopted. It is. In the following description, SGCK and BasicPhoto will be described as examples.

In each compression algorithm, first, whether the display color gamut is within the printer color gamut is determined. In the SGCK inside / outside determination, the intersection (C _I , J _I ) between the straight line connecting (0, J ₀ ) and the color point (C, J) to be determined and the outer shell of the display color gamut is within the printer color gamut. It is determined whether or not there is. On the other hand, in the basic photo inside / outside determination, the intersection (C _I ', J') between the straight line J = J 'including the color point (C', J ') to be determined and the outer shell of the display color gamut is within the printer color gamut. It is determined whether or not. Here, in SGCK, the intersection (C _I , J _I ) is the determination target for the inside / outside determination, and in Basic Photo, the intersection (C _I ′, J ′) is the determination target for the inside / outside determination, because the SGCK compression direction is (0 , J ₀ ) direction and the compression direction of BasicPhoto is the J = J ′ direction. For other compression algorithms, the inside / outside determination of the intersection of the straight line extending in the compression direction from the determination target color point and the display color gamut outer shell is performed. Color points for which the intersection point is determined to be within the printer color gamut in this inside / outside determination are mapped without changing the coordinate values.

On the other hand, in order to determine whether or not the intersection (C _I , J _I ) or the intersection (C _I ′, J ′) in the inside / outside determination is out of the printer color gamut is a color point to be compressed. The compression target determination is performed. The compression target determination is performed by determining whether or not the color point to be processed is inside a predetermined uncompressed region P1 in the printer color gamut. The uncompressed area P1 in this embodiment is an area in which the printer color gamut is compressed to 90% in the saturation direction, and the saturation in the uncompressed area P1 is 90% or less of the saturation in the printer color gamut outer shell. The display color gamut located in the uncompressed area P1 is mapped to the same coordinate value of the printer color gamut without being compressed. On the other hand, the display color gamut that protrudes outside the uncompressed area P1 is linearly compressed and associated with the compressed area P2. The compression area P2 is an area set in the vicinity of the outer shell of the printer color gamut, and in this embodiment, the remaining printer color gamut excluding the non-compression area P1 corresponds to the compression area P2.

  In the present embodiment, the non-compressed area P1 is described as 90% of the printer color gamut. However, the present invention is not limited to this percentage, and any percentage greater than 0 and less than 100 can be adopted as appropriate. Further, it is not always necessary to specify the same percentage over the entire lightness range, and the non-compressed region P1 is a non-compressed region P1 as long as it is a region that includes the outer shell of the printer color gamut and extends inside the printer color gamut. Various ranges can be set by changing the size of the compression region P1.

When the compression target determination is completed, in the first compression algorithm, linear mapping as shown in FIG. 7 is performed on the color point determined as the compression target. FIG. 7 is a diagram for explaining the first compression algorithm. The mapping shown in the figure can be expressed by, for example, the following formula (1).

Here, (C, J) is the coordinate of the color point before compression, and (C _new , J _new ) is the coordinate of the color point after compression. In addition, on the straight line connecting (0, J ₀ ) and the color point (C, J), the distance from (0, J ₀ ) to the color point is set to D ₁ , (0, J ₀ ) to the uncompressed region P 1. _D 2 the distance to the outer shell of, as (0, _{J 0)} distance from to the outer shell of the display color gamut of _{D 3, (0, J 0} ) D 4 the distance to the outer shell of the printer gamut from is there.

That is, in the above formula (1), the distance from the outer shell of the non-compressed region P1 to the outer shell of the display color gamut (C ₁ -D ₂ ) D ₃ -D ₂ ) and the ratio of the distance (D ₄ -D ₂ ) from the outer shell of the non-compressed region P1 to the outer shell of the printer color gamut, so that the outer shell of the non-compressed region P1 and the display color gamut The positional relationship of the color point between the outer shell and the outer shell of the non-compressed region P1 is normalized and reproduced between the outer shell of the printer color gamut. As a result of the above compression, the color point having a lightness higher than the peak value among the color points outside the printer color gamut is mapped to the compression region P2 inside the printer color gamut by compressing the lightness and the saturation. In addition, the compression algorithm of Formula (1) is an example, and it does not matter even if it maps not only with linear mapping but with a nonlinear algorithm.

On the other hand, in the second compression algorithm, linear mapping as shown in FIG. 8 is performed on the color points determined to be compressed. FIG. 8 is a diagram for explaining the second compression algorithm. The mapping shown in the figure can be expressed by the following equation (2), for example.

Here, (C, J) is the coordinate of the color point before compression, and (C _new , J _new ) is the coordinate of the color point after compression. Further, the saturation of the outer shell of the uncompressed region P1 at lightness J is C ₁ , the saturation of the outer shell of the printer color gamut at lightness J is C ₂ , and the saturation of the outer shell of the display color gamut at lightness J is C _3. It is as. In the equation (2), the distance (C ₃ − from the outer shell of the non-compressed region P1 to the outer shell of the display color gamut with respect to the distance (C−C ₁ ) from the outer shell of the non-compressed region P1 to the color point. the C ₁₎ and multiplying the ratio of the distance from the outer shell of the uncompressed region P1 to the outer shell of the printer gamut (C 2 _{-C 1),} the outer shell and the outer shell of the display color gamut of the uncompressed region P1 The positional relationship between the color points is normalized and reproduced between the outer shell of the non-compressed region P1 and the outer shell of the printer color gamut. As a result of the above compression, among the color points outside the printer color gamut, color points having a lightness lower than the peak value are mapped to the compression region P2 within the printer color gamut by compressing the saturation while maintaining the lightness. . In addition, the compression algorithm of Formula (2) is an example, and it does not matter even if it maps not only with linear mapping but with a nonlinear algorithm.

Or more at the first compression algorithm described, as increased compression degree of brightness away from the peak value, the compression degree of brightness closer to the peak value is low, the brightness of the compression ratio in the peak J ₀ 0 It becomes. Of course, in the second compression algorithm, the lightness is maintained regardless of the lightness. That is, at the peak value, the compression rate and the compression direction of the first compression algorithm coincide with the compression rate and the compression direction of the second compression algorithm. Therefore, while adopting different compression algorithms across the peak value, continuity in the peak value is ensured between the mapped color points, and brightness reversal in the vicinity of the peak value is also prevented. Become. By preventing the reversal of lightness, it is possible to realize color gamut conversion that realizes a print result that gives a natural impression to human eyes that are more sensitive to lightness than saturation.

  FIG. 9 is a diagram for explaining the effect of using different compression algorithms with the above-described peak value as a boundary. This figure shows the gradation of green, and image data in which the brightness of green is continuously changed from the lower left to the upper right of the circle is mapped and printed by each compression algorithm. 9A is a printing result when mapping is performed with BasicPhoto over the entire brightness range, FIG. 9B is a printing result when mapping is performed with SGCK over the entire brightness range, and FIG. Print results when mapping is performed using a compression algorithm are shown. In the printing result of FIG. 9A mapped with BasicPhoto, the high gradation in the upper right is whitish and looks bad, and in the printing result of FIG. 9B mapped with SGCK, the brightness changes greatly in the dark part. Therefore, the gradation in the dark part is poor and the black crescent moon is clearly visible. On the other hand, in FIG. 9C mapped by performing the color gamut conversion processing of the present embodiment, it can be seen that the gradation is good in the high saturation and the gradation is maintained in the dark portion.

FIG. 10 is a flowchart of the color gamut conversion process performed using the equations (1) and (2). In the figure, first, the hue angle θ to be processed is determined in step S300, and initial coordinates (C, J) at the hue angle θ are determined in step S305. And it obtains the peak value J ₀ of the printer gamut in the hue angle θ at step S310, the in step S315, performs outside judgment of the printer gamut to the color point of the initial coordinate. As a result of the inside / outside determination, if the inside of the printer color gamut is determined, the color is mapped to the color point of the printer color gamut having the same coordinate value as (C, J) in step S320. On the other hand, if it is determined that it is outside the printer color gamut, it is determined which compression algorithm to use in step S325 based on the peak value acquired in step S310. In step S330, the coordinates (C _new , J _new ) of the compression destination are determined according to the determined compression algorithm, and mapped to these coordinates.

In this way, when the mapping of the coordinates (C, J) is completed in steps S315 and S330, it is determined in step S335 whether or not the calculation of (C _new , J _new ) for all coordinates has been completed. Here, since this color gamut conversion process is a process as a process for creating an LUT, even if it is determined whether or not the calculation has been completed for all coordinates, it is necessary and sufficient to create the LUT. It is sufficient to determine whether or not the calculation for the coordinates has been completed. The calculation is performed only on the integer coordinate value, or the calculation is performed only for the lattice point in consideration of the lattice point having a predetermined pitch in the J ^* a ^* b ^* space. Various modes such as making it possible can be adopted.

  If it is not determined in step S335 that the calculation has been completed for all coordinates, the next calculation candidate coordinate (C, L) is set in step S345, and the processes in and after step S315 are repeated. If it is determined in step S335 that the calculation has been completed for all coordinates, it is determined in step S340 whether or not the gamut conversion processing has been completed for all hues, and it is determined that the gamut conversion processing has been completed for all hues. Steps S300 and after are repeated until

  Note that the color gamut conversion processing of the present invention is applicable not only when statically creating an LUT as described above but also when dynamically generating a mapping destination for an input sRGB value. . When continuously processing the color points of each hue angle as in the LUT creation, it is more efficient to determine the peak value in step S310 before the color gamut inside / outside determination in step S320 as in the flowchart described above. However, when the hue angle of the color point input in the dynamic mapping process changes randomly, it is more efficient to perform the peak value determination in step S310 after the color gamut inside / outside determination in step S320. Get better.

Incidentally, as a boundary the peak value J ₀ of the printer gamut in the embodiment described above, the high brightness side executes the color gamut conversion at a first compression algorithm, the low-brightness side in the second compression algorithm Although color gamut conversion is performed, in addition to switching the compression algorithm depending on whether the peak value is higher or lower than the peak value, for example, by switching the compression algorithm according to the similarity of the outer shell shape Also good. For example, if a color gamut having an upwardly convex outer shell shape is mapped to a color gamut having a downwardly convex outer shell shape, changes in lightness and saturation are emphasized, resulting in a sharp gradation. A big influence. In view of these circumstances, a brightness maintaining compression algorithm that maintains gradation is used when the bulge direction of the outer shell shape is different. Adopt a compression algorithm to compress the degree. Then, the gradation is appropriately maintained according to the color gamut shape.

(4-2) Modification of color gamut conversion process:
Next, a modification of the above-described color gamut conversion process will be described. In this modification, according to the degree of similarity of the color gamut shape between the printer color gamut and the display color gamut, the entire lightness is switched without switching the compression algorithm with the peak value or switching the compression algorithm as in the above-described embodiment. Select whether to apply a lightness-maintaining compression algorithm over a range. The determination of the similarity degree, utilizing the positional relationship between the peak value L _cusp 'of peak L _cusp and the display color gamut of the printer gamut. In other words, when the peak values are close to each other between the printer gamut and the display gamut, the gamut shapes are considered to be similar, and in mapping the color points outside the printer gamut in the display gamut into the printer gamut, The mapping is performed while maintaining the lightness substantially. Thus, by switching the compression algorithm based on the relationship between the peak value of the printer color gamut and the peak value of the display color gamut, the gradation after mapping is more appropriately maintained.

FIG. 11 shows a state in which the display color gamut J ^* a ^* b ^* and the printer color gamut P are cut along the ab plane in the J ^* a ^* b ^* space. (A) is a state cut at a green hue angle, (b) is a state cut at a magenta hue angle, and (c) is a state cut at a blue hue angle. In the figure, the horizontal axis represents saturation C, and the vertical axis represents lightness J. The peak value in each hue of the printer gamut has a different brightness value for each hue angle, with high brightness in green, medium brightness in magenta, and low in blue. Present in lightness. Therefore, if mapping between color gamuts where the peak value J ₀ and the peak value J ₀ ′ are close to each other, even if the color gamut conversion processing is performed with the lightness maintaining compression algorithm over the entire lightness range, If the mapping is between color gamuts where the peak value J ₀ and the peak value J ₀ ′ are separated, the brightness is shifted in addition to the saturation shift. Is preferable from the viewpoint of color reproduction.

FIG. 12 is a diagram for explaining how to switch the compression algorithm in the color gamut conversion processing of the present modification. As shown in the figure, in the color gamut conversion process of the present modification, first, the necessity of switching the compression algorithm is determined based on the peak value of each color gamut using the following equation (3).

Here J ₀ is the peak value of the printer gamut, J ₁ is the peak value of the display color gamut, J _w white point of the printer gamut (display color gamut), a. According to this equation (3), half of the white point interval of peak J ₀ as a threshold J _th, compression algorithms based on the magnitude relation between the difference J _sub and the threshold J _th between peak switching It is determined whether or not to perform. That is, when the difference between the peak values is larger than the threshold value (J _sub > J _th ), it is determined that the printer color gamut and the display color gamut are not similar, and the peak is the same as in the above-described embodiment. performing color gamut conversion processing is switched to a first compression algorithm to a value J ₀ at the boundary with the second compression algorithm. On the other hand, if the difference between the peak values is less than or equal to the threshold value (J _sub ≦ J _th ), it is determined that the printer color gamut and the display color gamut are similar, and for example, the lightness is maintained over the entire lightness range. In addition, the color gamut conversion process is executed by the second compression algorithm for compressing the saturation.

The compression direction of the first compression algorithm in the embodiment described above is the direction from the color point (C, J) to be processed toward the peak value (0, J _0p ) of the printer color gamut. In the example, a change corresponding to the difference J _th between the peak values may be added in this direction. Specifically, it is conceivable to set the compression direction in consideration of the angle according to the difference J _th as shown in the following formula (4).

Here, the vector n ₁ is a vector from the color point (C, L) to the peak value (0, L ₀ ) of the printer color gamut, and the vector n ₂ is from the peak value (0, L ₀ ) of the display. A vector toward the peak value (0, L ₀ ′) of the printer color gamut, and n _comp is a vector representing a compression direction in consideration of an angle corresponding to the difference L _th . By this equation (4), the n _comp direction obtained by multiplying the vector n ₁ by multiplying the vector n ₂ by a predetermined coefficient α is determined, and by compressing in the n _comp direction, the peak value of the lightness compression degree in the compression is determined. The difference between them will be reflected. That is, mapping can be performed in which the degree of lightness compression increases as the distance between peak values increases, and the degree of lightness compression decreases as the distance between peak values decreases. The embodiments lightness compression ratio in the peak value L ₀ of the printer gamut in the present modification becomes 0, the continuity with the compression direction and the compression ratio in peak L ₀ is maintained in the aforementioned It is the same.

FIG. 13 is a flowchart for performing a color gamut conversion process while selecting a mapping algorithm based on the equation (3). When the process is started, first, a hue angle θ at which a color point to be processed exists is determined in step S400, and initial coordinates (C, J) at the hue angle θ are determined in step S405. And it obtains the peak value J ₀ of the printer gamut in the hue angle θ at step S410, in step S415, performs outside judgment of the printer gamut to the color point of the initial coordinate. As a result of the inside / outside determination, if the inside of the printer color gamut is determined, the color is mapped to the color point of the printer color gamut having the same coordinate value as (C, J) in step S420. On the other hand, if it is determined that it is outside the printer color gamut, the peak value J ₀ of the printer color gamut and the peak value J ₀ ′ of the display at the hue angle θ are determined in step S425, and the difference between these peak values. J _sub is compared with the threshold value J _th . If the difference L _sub is smaller than L _{th as} a result of the comparison, it is determined in step S430 that the algorithm is not switched, and the process proceeds to step S435, where only the algorithm that maintains the lightness without switching the compression algorithm is used. Decide to do the mapping. On the other hand, when it is determined that the difference L _sub is larger than L _{th as} a result of the comparison, the process proceeds to step S440, and it is determined to perform mapping while switching the compression algorithm. Then, mapping is performed using the determined method. Then, in accordance with the determined compression algorithm, the coordinates (C _new , J _new ) of the compression destination are determined in step S345 and mapped to these coordinates.

When the mapping of the coordinates (C, J) is completed in steps S420 and S445 in this way, it is determined in step S450 whether or not the calculation of (C _new , J _new ) for all the coordinates has been completed. If it is not determined in step S450 that the calculation has been completed for all coordinates, the next calculation candidate coordinates (C, L) are set in step S455, and the processes in and after step S415 are repeated. If it is determined in step S450 that the calculation has been completed for all coordinates, it is determined in step S460 whether or not the gamut conversion processing has been completed for all hues, and it is determined that the gamut conversion processing has been completed for all hues. Step S400 and subsequent steps are repeated until

(5) Smoothing processing of peak value fluctuation between hues:
In the above color gamut conversion processing, processing using the peak values of the printer color gamut and display color gamut is performed. FIG. 14 is a graph in which the peak value in each hue of the printer color gamut is plotted with respect to the peak value. In the figure, portions where the amplitude occurs in small increments are observed at hue angles of about 110 °, 130 °, and 260 °. In a portion where a small amplitude is generated, when the above-described color gamut conversion process is performed, inversion of lightness occurs before and after the amplitude, and a pseudo contour may occur. In order to prevent the occurrence of such a pseudo contour, it is preferable to smooth the gradation between hues. Hereinafter, processing for smoothing peak value fluctuations between hues for preventing the occurrence of pseudo contour will be described.

In the smoothing of the peak value variation between hues in the present embodiment, the peak value of the hue having the maximum saturation is used as a reference point for smoothing. In general, the primary color has a high color expression capability, so it is easy to obtain the maximum saturation. Therefore, first, the peak value of the primary color in the printer color gamut (in this embodiment, RGBCMY is assumed as the primary color of the additive color mixture and subtractive color mixture) is adopted as a smoothing reference point, and the following formula (5 ) To smooth the peak values between adjacent primary colors.

Here, L ₀ (θ) is a peak value in a predetermined hue θ (θ _B ? Θ? Θ _M ), L _0M is a peak value in a magenta hue θ _M , and L _0B is a peak value in a blue hue θ _B. The value Rate is a coefficient for linear interpolation between the peak values of adjacent primary colors. In addition, although the formula for smoothing the hue angle of blue to magenta is illustrated in the formula (5), the peak value between other primary colors can be smoothed by the same formula. FIG. 15 is a graph obtained by approximating the graph of FIG. 14 using the peak value of the primary color. It can be seen that the portion that was oscillating in small increments was smoothed. If interpolation is performed between the peak values of the default primary color in this way, peak value calculation processing and interpolation processing become very simple.

By the way, there is a case where the primary color does not take the maximum saturation, for example, as in the JCh color space. In such a case, first, a reference point having the maximum saturation is detected. That is, instead of the primary color, the peak value of a hue whose inclination of the peak value greatly varies is used as a reference point. That is, the hue angle θ that the degree of change in peak J ₀ to the change in hue varies greatly (hereinafter, referred to as the slope change point I.) To a plurality of points detected. Then, by interpolating between the peak values L ₀ (I) at each gradient fluctuation point I with a predetermined interpolation formula, the change in the peak value L ₀ with respect to the change in hue is smoothed.

In order to detect the gradient fluctuation point I, first, the peak value J ₀ of the color gamut obtained by cutting the printer color gamut at a predetermined hue angle is sequentially obtained while changing the hue angle θ by a predetermined amount, and the peak value J for the hue is obtained. ₀ fluctuation data is acquired. Then setting a predetermined hue range R, the degree of change of the slope of the peak value J ₀ at a predetermined hue range R within a predetermined hue range R of the acquired variation data, by a predetermined amount a predetermined hue range R Calculation is performed over the entire hue while shifting. The minimum value of the predetermined hue range R set at this time needs to be set wider than the small amplitude described above. Then, for example, twelve hue angles θ _{1 to} θ ₁₂ are selected in descending order of variation, and set as gradient variation points I _{1 to} I _{12 in} ascending order of hue angles. As a more specific method of detecting the gradient fluctuation point, it may be a point where the sign of the gradient changes before and after the gradient fluctuation point, such as a, b, c in FIG. It is possible that the slope sign itself does not change as in 14 d, but the slope changes relatively large. In practice, the portion that is likely to be set at the above-described gradient variation point is often an extreme value or an inflection point.

Then, the peak value of the determined gradient change point is interpolated by the following interpolation formula (6).

In the equation (6), J ₀ (θ) of θ _n ? Θ? Θ _{n + 1} can be obtained. _Here, J 0 _{(I n)} is the peak value of the gradient change point _{I n, J 0} (θ) is the peak value of the hue angle theta, Rate 'is linear interpolation between peak values of adjacent reference points It is a coefficient. According to the above formula (6), the peak value between the peak values of each gradient fluctuation point is smoothed, and the amplitude can be eliminated in small increments. The interpolation between the peak values of the reference points is not limited to linear interpolation, and high-order interpolation such as cubic interpolation may be used.

FIG. 16 is a flowchart showing a flow of processing for smoothing peak value fluctuations between hues. In the present embodiment, in step S600, the peak value at each hue angle is obtained for every hue angle. In step S605, the inclination of the acquired peak value with respect to the hue change is detected, and the hue angles θ _{1 to} θ ₁₂ having a large inclination change are determined as the gradient change points I _{1 to} I ₁₂ in ascending order of the hue angle. In step S610, the peak values L (I ₁ ) to L (I ₁₂ ) at the gradient fluctuation points I _{1 to} I ₁₂ are linearly interpolated using Expression (6). If the above-described color gamut conversion processing is performed using the peak values of the hues obtained as described above, the mapping vector direction between the hues is stabilized, and the small amplitude resulting from the peak value change with respect to the hue change Is suppressed, and pseudo contour is not generated.

  Further, in smoothing the peak value fluctuation between hues, in addition to the six primary colors described above, an intermediate value of each primary color is added to the reference point, or a reference point between the detected gradient change points I is added. If a plurality of values are added, the accuracy of approximation of the interpolated peak value to the actual peak value can be expected. Further, as a method of determining the gradient fluctuation point I described above, for example, only a portion where the gradient changes positively or negatively may be adopted as the gradient fluctuation point I. This is because when the sign of the slope does not change, approximation can be performed by interpolation or the like, and it is assumed that there is no significant deviation from the actual peak value. Of course, the interpolation is not limited to linear interpolation between peak values, but it is of course possible to perform interpolation in which each peak value is approximated nonlinearly by a cubic interpolation equation or the like. When cubic interpolation is used, approximation accuracy is improved when the fluctuation of the peak value at the reference point is moderate.

Furthermore, the above-described embodiments may be combined as a suitable modification when it is unclear whether the primary color is a gradient change point. That is, while setting the primary color RGBCMY as a temporary gradient change point, the peak value J ₀ of each hue included in the predetermined hue range R substantially centered on the primary color is sequentially obtained, and the obtained fluctuation data The degree of inclination variation is calculated. If the degree of change exceeds a predetermined reference, the primary color RGBCMY is adopted as the true gradient change point. On the other hand, if the degree of variation does not exceed a predetermined standard, there is a possibility that another true gradient variation point may exist, so the search for the hue in which the inclination of the peak value varies greatly is performed. . In this way, it is possible to reliably adopt the true gradient change point by performing the process of confirming whether the primary color is really the gradient change point, while assuming that the primary color is likely to be the gradient change point. Furthermore, when determining whether or not the primary color is a gradient change point, the amount of calculation can be minimized by making only the predetermined hue range R of each primary color a check target. In general, assuming that the primary color is often a gradient change point, the case where the primary color is not a gradient change point is an exception process.

The above-described processing for smoothing the peak value between hues is preferably applied to a printer color gamut having uneven gamuts, particularly a color gamut such as photomat paper, but of course may be applied to the display color gamut. Absent. In particular, if the gradient change point and the primary color match, instead of sequentially calculating the peak value of each hue angle, the peak value of each hue angle is predicted by the hue gradation smoothing process. It is also possible to improve the processing speed.
(6) Speeding up the reverse conversion process:
Incidentally, the second color conversion table used in the reverse direction conversion process is a forward multidimensional LUT for converting RGB values into XYZ values, and is arranged in ascending order of RGB values as shown in FIG. Therefore, there is no regularity in the arrangement order of the XYZ values, and in order to search for RGB values corresponding to the XYZ values input in the reverse conversion process, the reverse lookup in the forward multidimensional LUT, that is, the XYZ values are all set. Processing to retrieve and approximate XYZ values has been required. However, since reverse lookup causes an increase in processing time, an improvement in processing speed has been desired.

  Therefore, in the present embodiment, the XYZ value calculation module M2 includes a prediction module M21 and a search module M22. The prediction module M21 predicts the appearance position of the RGB value corresponding to the input XYZ value in the sRGB color space, and determines the optimum search order. The search module M22 searches for approximate XYZ values according to the search order determined by the prediction module M21. Hereinafter, the prediction method of the prediction module M21 and the search order determined based on the prediction result will be described.

  FIG. 18 is a schematic diagram for explaining a prediction method of the prediction module M21. For the sake of simplicity, the drawing shows the RGB color space in two dimensions cut along an arbitrary GB plane. However, in actuality, the number of dimensions according to the order of the color space is required, as in this embodiment. Needless to say, prediction is performed in a three-dimensional space in the sRGB color space. In the figure, the color gamut X of the printer 40 in the sRGB color space is divided into a plurality of regions. The division logic is an analogy of bifurcated division in the binary search method. The entire color gamut X is equally divided into four areas C1 to C4, and each of the areas C1 to C4 is equally divided into four areas B1 to B1. B4 is set, and regions A1 to A4 are set by dividing each of regions B1 to B4 into four equal parts. In FIG. 18, the areas B1 to B4 and the areas A1 to A4 divided into four areas are set by taking the area C1 and the area B4 as an example, but the areas similar to the areas B1 to B4 are also set in the areas C2 to C4. In the areas B1 to B4, the same area setting as that of the areas A1 to A4 is performed. In the present embodiment, the number of divisions at each stage is set to four, and three stages of division are performed. However, the number of divisions at each stage is arbitrary, and the number of divisions can be arbitrarily selected.

  By the way, the XYZ values converted in the reverse conversion process are originally RGB values of the display 20 and are converted in the order in which the RGB values of the display 20 are input. That is, the order in which the color gamut conversion process is performed when creating the LUT 15b is the ascending order of the RGB values of the display 20. Therefore, the prediction module M21 stores the previous RGB value converted from the XYZ value converted immediately before, and uses it for the prediction of the conversion destination of the next input XYZ value. That is, when (X1, Y1, Z1) that has been subjected to the color gamut conversion processing immediately before is converted into Px (R1, G1, B1) in the color gamut of the printer 40, the area A2 including this Px is stored. . Then, when searching for (R2, G2, B2) which is the conversion destination of the next input (X2, Y2, Z2), the search module M22 is first searched for the area A2.

  Further, the RGB data is generated as a data string in which coordinate values are arranged in the order of RGB from the top as shown in FIG. Accordingly, when data conversion is performed in the order of generation, first, the coordinate value of B arranged in the least significant bit of (R, G, B) is incremented bit by bit. When B reaches 255, the G coordinate value is incremented by 1 bit, and then the B coordinate value increases from 0 to 255 again. When the G coordinate value reaches 255 bits, the R coordinate value is incremented by 1 bit, and the B coordinate value is increased from 0 again. That is, the RGB value processed in the color conversion process should be an RGB value shifted by 1 bit in the direction of increasing B in the color gamut of the display 20 from the RGB value processed immediately before. Furthermore, the color space after the color gamut conversion processing is limited to the printer color gamut, but is an sRGB color space. Therefore, the RGB value in the printer color gamut behaves similarly to the variation of the RGB value in the color gamut of the display 20. It is estimated that Based on this estimation, when the search module M22 cannot find (X2, Y2, Z2) in the XYZ values corresponding to the RGB values in the area A2, the search module M22 searches by predicting the fluctuation direction of the RGB data. Enlarge the area.

  The search order will be described in more detail with reference to FIGS. FIG. 20 is a diagram for explaining the search order in the RGB color space, and FIG. 21 is a flowchart for selecting the search order. In FIG. 20, the region P of the RGB color space is divided into eight equal parts. In the figure, the coordinates of the conversion destination of (X1, Y1, Z1) converted by the color gamut conversion processing immediately before (X2, Y2, Z2) to be converted are Px (R1, G1, B1). , Px is an area P0. In such an RGB color space, the first search target area is a region P0 where Px exists. Then, while sequentially acquiring the conversion destination of the RGB value of the region P0 from the forward direction LUT, the coordinates (proximity XYZ value) close to (X2, Y2, Z2) are searched from the acquired XYZ values. The determination as to whether or not they are close may be made, for example, by determining that an object closer than a predetermined threshold is close, and the threshold may be a lattice point interval or the like.

  When a predetermined number of proximity XYZ values are found, the RGB values corresponding to the found proximity XYZ values are acquired from the forward conversion LUT, and the transformation coordinates Py (R2, G2, B2) are determined by performing an interpolation operation. . The predetermined number is, for example, 6 points if hexahedral interpolation is performed, and 4 points if tetrahedral interpolation is performed. On the other hand, when the proximity XYZ value cannot be found, the region P4 in the direction in which the coordinate value of B increases is determined as the next search target area. Thereafter, when the proximity XYZ value cannot be found also in the region P4, the region is changed in the order of P2, P6, P1, P5, P3, and P7, and each search is performed. In addition, the search order when the first search target area is P1 is P1 → P5 → P3 → P7 → P0 → P4 → P2 → P6, and the search order when the first search target area is P4. Is P7-> P3-> P5-> P1-> P6-> P1-> P4-> P0.

  Based on the search order determined as described above, the flow of the backward conversion process executed by the XYZ value calculation module M2 will be described with reference to FIG. Note that the description of this process is based on the two-dimensional region drawing of FIG. 18 for the sake of simplicity. When the process is started, the XYZ values first input from the CA conversion module M5 are acquired as initial data in step S500. In step S505, an XYZ value close to the initial data of the XYZ value is searched from the forward LUT, and an interpolation operation based on the conversion destination RGB value of the close XYZ value acquired in all searches is executed in step S510. The RGB value obtained by the interpolation calculation is registered in the LUT 15b as the XYZ value conversion destination, and is temporarily stored as Px in step S515. In step S512, it is determined whether or not data input in a series of color gamut conversion processes is completed before execution of step S515. If completed, the process ends. If not, the process ends. Proceeding to S515 and S520, the next XYZ value is acquired from the CA conversion module M5.

  When the process proceeds to step S525, the narrowest area A among the areas including Px in FIG. 18 is set as the prediction area, and the adjacent XYZ values of the XYZ values acquired in step S525 are searched from the forward LUT in the area A. To do. If a proximity XYZ value is found in the region A as a result of the search, the process returns from step S530 to step S510, and an interpolation operation based on the converted RGB value of the proximity XYZ value found is executed. On the other hand, if the proximity XYZ value is not found in the region A even if the region A is searched, the process proceeds to step S535, the search region is expanded, and the region B is searched. Also in the search for the area B, the search is executed in order from the area located in the B-axis direction from the area A where Px exists. When the proximity XYZ value is found in the region B, the process returns to step S510 to execute an interpolation operation based on the converted RGB value of the found proximity XYZ value. On the other hand, if the proximity XYZ value is not found even after searching for the area B, the process proceeds to step S545, where the search area is expanded and the area C is searched. When the proximity XYZ value is found in the area C, the process returns to step S510 to execute an interpolation calculation based on the converted RGB value of the found proximity XYZ value. On the other hand, if the proximity XYZ value is not found even after searching the area C, the process returns to step S505 as an exception process, and the proximity XYZ value is searched by the all search process.

(7) Color gamut inside / outside determination processing:
As described above, in the color gamut conversion processing executed by the color gamut conversion module 24, the color gamut conversion is performed so that the printer color gamut and the display color gamut coincide. However, near the surface of the color gamut, the color gamut setting method performed by the color gamut conversion module and the color gamut setting method performed by the XYZ value calculation module are different from each other, or due to errors, calculation errors, etc. It may be mapped outside the color gamut. Since a color point that has been mapped outside the color gamut cannot be reversed, it is necessary to change (remapping) the mapping destination within the color gamut. More specifically, the gamut inside / outside determination processing is executed for the XYZ values calculated in the CA conversion module M4, and the above-described backward conversion processing is executed for the coordinate values determined to be within the gamut. The coordinate value determined to be out of the gamut is remapped to the nearest point in the color gamut. Hereinafter, the color gamut inside / outside determination processing will be described.

FIG. 23 is a diagram for explaining the outline of the inside / outside determination processing according to this embodiment. The figure is a graph showing by cutting the J ^* a ^* b ^* color space prescribed ab plane, Px is the coordinate value to be determined, O is a point on the gray axis, P is OP vector and gamut The intersection with the shell. Further, as shown in FIG. 24, the color gamut of this embodiment is obtained by further dividing a hexahedron formed by connecting the lattice points of the LUT into six tetrahedrons. For example, if the LUT stores RGB data (729 colors) with 9 slices of the range 0 ≦ R, G, B ≦ 255 and data describing the corresponding XYZ values, connect these data. There are 512 hexahedrons, and each hexahedron is further divided into six tetrahedrons. That is, the outer shell of the color gamut is formed by connecting triangular planes to form a so-called polygon. Hereinafter, among the triangular planes that form the tetrahedron, the surface that forms the outer shell of the color gamut will be referred to as a mesh. In the inside / outside determination process, a mesh that intersects the vector OPx is searched from among the meshes that form the outer shell, and an intersection point P between the found mesh and the vector OPx is calculated. Then, | OP | (= A) and | OPx | (= B) are compared. When B> A, Px is determined to be out of gamut and remapping is performed. When A> B, Px is gamut. The above-described color gamut conversion process is continued.

With reference to FIG. 25, the search method of the mesh which cross | intersects OPx is demonstrated. FIG. 25 is a flowchart of the inside / outside determination process. When the process is started, initial data of XYZ values is acquired in step S700. In step S705, a search is made for a mesh that intersects the vector OPx using the acquired XYZ value as Px. This search target is all meshes constituting the outer shell of the color gamut of the printer 40. Whether or not the mesh and the vector OPx intersect can be determined by, for example, the following method. As shown in FIG. 26, each vertex of the mesh trihedron is represented as ABC, the input output device color data is represented as P, and can be represented by the following equation (7).

If P is inside the triangle ABC, in equation (7) above
Is established. That is, if equation (8) holds in equation (7), it is determined that vector OPx intersects with the mesh, and if not established, vector OPx can be determined not to intersect with the mesh.

  In the subsequent step S710, the mesh determined in step S705 is stored. In step S715, the intersection of the hit-determined mesh and the vector OPx is obtained, and the inside / outside determination of the processing target coordinates is performed. If the inside / outside determination determines that the color gamut is inside, the process proceeds to step S725 to continue the LUT creation process. If it is determined that the color gamut is outside, the process proceeds to step S720 to perform remapping. The remapping destination is the mesh closest to the mesh determined in step S705 and the mesh adjacent to the mesh. When the remapping is completed, the process proceeds to step S725, and the LUT creation process is continued. In step S730, it is determined whether the internal / external determination has been completed for all data. If all the data is finished, the inside / outside determination process is finished. If all the data is not finished, the process proceeds to step S735, and the next XYZ value is acquired. In step S740, using the mesh stored in step S710, a hit determination is performed using the stored mesh and its adjacent mesh as search targets for the hit determination. If there is a hit determination in step S740, the process proceeds from step S745 to step S710 to store the hit mesh. On the other hand, if there is no hit determination, the process proceeds from step S745 to step S705 to search for all meshes. Become.

Note that the color space for determining the hit between the XYZ value and the triangular plane and calculating the intersection may be performed in the XYZ color space or in the J ^* a ^* b ^* color space. When using the J ^* a ^* b ^* color space, the J ^* a ^* b ^* value corresponding to the acquired XYZ value and the outer mesh of the RGB color space are used to generate the J ^* a ^* b using the forward LUT. ^{* It} is only necessary to execute a hit determination or an intersection calculation with the one converted into the color space. Further, when a polar coordinate system such as JCh is adopted as a space for performing the color gamut conversion processing in the color gamut conversion module M5 instead of an orthogonal coordinate system such as a J ^* a ^* b ^* color space, the distance to Px and the intersection P Is easily calculated.

  Further, as a suitable example when performing a full search, it is conceivable to sort the mesh data according to lightness and hue angle. That is, when an XYZ value is input, brightness and hue angle are calculated from the XYZ value using a known conversion formula, and a mesh having a high possibility of having an intersection with a color point to be subjected to inside / outside determination processing is narrowed down. I can do it. Therefore, when performing a full search, the sum of meshes for performing the hit determination can be narrowed down, so that the processing can be speeded up.

  Further, as a more specific example of the present embodiment, the hit determination and intersection calculation may be executed by an external API optimized for hit determination and intersection coordinate calculation between an arbitrary line and an arbitrary plane. . For example, DirectX can be considered as an external API. DirectX is executed using the graphics board GPU (Graphics Processing Unit) and high-speed graphics memory, so it can be processed at high speed. It has a function to execute at high speed. Therefore, by giving the vector OPx and the coordinate value of each vertex of the mesh to DirectX and executing a predetermined calculation, the hit determination and the calculation of the intersection coordinates can be performed at high speed.

In order to suppress the occurrence frequency of remapping, the following method may be used.
First, the same method is adopted as the region setting method used in determining the color gamut performed by the color gamut mapping module M5 and the region setting method including the printer color gamut performed when creating the LUT. Then, since the deviation of the region outer shell is eliminated, the number of occurrences of the remapping described above is reduced.
Second, in the area setting of the printer color gamut, an area formed by a method of faithfully connecting the grid points of the second color conversion table is set. This is because the color gamut of the printer has irregularities in the outer shell, and for example, if a three-dimensional convex hull is used, the irregularities are crushed and the color accuracy deteriorates, resulting in a difference from the actual color gamut shape. The number of occurrences of remapping is reduced by preventing the deterioration of color accuracy.
Thirdly, as the compression algorithm in the color gamut mapping module M5, for example, the above-described SGCK or BasicPhoto is adopted. Since these compression algorithms map in a certain direction without changing the hue, there is no lightness or re-switching in the data before and after mapping. Therefore, the coordinate value converted to the color gamut of the printer 40 is converted to the vicinity of the coordinate value converted immediately before, and the prediction hit rate in the backward conversion process described above is improved.
Fourth, in performing color gamut conversion, the above-described processing for smoothing the peak value fluctuation is executed. Then, the mapping data changes gradually, and the data after mapping is generated at a constant interval. Therefore, the coordinate value converted to the color gamut of the printer 40 is converted to the vicinity of the coordinate value converted immediately before, and the prediction hit rate in the backward conversion process described above is improved.

(8) Summary:
As described above, the LUT for converting the device color of the display 20 to the device color of the printer 40 is created, the color gamut of the printer 40 is formed by polygons, and the color conversion table is created from the gray axis of the color gamut. Means to determine a hit between the half line extending to the coordinates after conversion and the polygonal plane, calculate an intersection of the polygonal plane determined to be hit, a distance from the reference point to the intersection, and The distance from the reference point to the coordinate after mapping is compared to determine whether the coordinate after conversion by the color conversion table creation means is within the color gamut. Remaps within the color gamut. In the hit determination, first, a polygon plane that is determined to have an intersection with respect to the color-converted coordinates and a polygon plane that surrounds the polygon plane are determined. Therefore, it is possible to provide an image processing apparatus capable of speeding up the creation of the color conversion table.

  Note that the present invention is not limited to the above-described embodiments and modifications, and the structures disclosed in the above-described embodiments and modifications are mutually replaced, the combinations are changed, the known technique, and the above-described implementations. Configurations in which the configurations disclosed in the embodiments and modifications are mutually replaced or the combinations are changed are also included.

It is a block diagram which shows the hardware constitutions of the computer which implement | achieves the image processing apparatus concerning embodiment of this invention. It is a block diagram which shows the software configuration of a computer. It is an example of the screen which performs the setting process of color gamut adjustment. It is a conceptual diagram of a color conversion module. It is a flowchart of LUT creation work. FIG. 6 is a schematic diagram illustrating a state where a display color gamut and a printer color gamut are cut at a predetermined hue angle θ in a J ^* a ^* b ^* space. It is a figure explaining a 1st compression algorithm. It is a figure explaining a 2nd compression algorithm. It is a figure explaining the effect by using properly a compression algorithm on the boundary of a peak value. It is a flowchart of a color gamut conversion process. This shows a state where the display color gamut J ^* a ^* b ^* and the printer color gamut P are cut along the ab plane in the J ^* a ^* b ^* space. It is a figure explaining the switching method of the compression algorithm in the color gamut conversion process of a modification. It is a flowchart which performs a color gamut conversion process, selecting a mapping algorithm. It is the graph which plotted the peak value in each hue of a printer color gamut about the peak value. It is the graph which approximated the graph of FIG. 14 using the peak value of the primary color. It is a flowchart which shows the flow of the process which smoothes the peak value fluctuation | variation between hues. It is an example of the forward multidimensional LUT used in a reverse direction conversion process. It is a schematic diagram explaining the prediction method of a prediction module. This is the general memory storage order of RGB data. It is a figure explaining the search order in RGB color space. It is a flowchart which selects search order. It is a flowchart of a reverse direction conversion process. It is a figure explaining the outline of an inside / outside determination process. It is a figure which shows the example which divides | segments a hexahedron into six tetrahedrons. It is a flowchart of a hit determination process with a mesh. It is a figure explaining the hit determination processing with a mesh.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Control part, 10a ... CPU, 10b ... RAM, 11 ... ROM, 13 ... USB I / F, 14 ... Video I / F, 15 ... HDD, 15a ... Image data, 15b ... LUT, 16 ... Bus, 17 ... Operation Input device I / F, 20 ... display, 21 ... printer driver, 21a ... image data acquisition module, 21b ... color conversion module, 21c ... halftone processing module, 21d ... print data generation module, 22 ... input device driver for operation (DRV), 23 ... display driver (DRV), 24 ... color gamut conversion module, 30 ... operation input device, 30a ... keyboard, 30b ... mouse, 40 ... printer, 100 ... computer, M1 ... XYZ value calculation module, M2 ... XYZ value calculation module, M3 ... CA conversion module, M4 ... CA conversion module Yuru, M5 ... gamut mapping module

Claims (8)

A correspondence relationship between the first color space of the first device-dependent color and the second color space of the second device-dependent color is set, and the first color in the first color space is determined based on the correspondence relationship. An image processing method for converting to a second color in a second color space,
A color conversion table setting step for setting a color conversion table to be referred to when converting the first color into the second color based on a predetermined conversion rule;
A shell setting step of setting a shell of the second color space by combining a plurality of polygonal planes;
A detecting step of detecting a polygonal plane having an intersection point with a half line from a predetermined point in the second color space to a color coordinate of the color converted based on the color conversion table;
An inside / outside determination step of comparing the intersection point with the color coordinates and determining whether the color coordinates are within the outer shell;
A mapping step of associating the color coordinates in the outer shell when the color coordinates are determined not to be in the outer shell in the inner / outer determination step;
Comprising
In the detection step, first, a polygon plane detected immediately before and a surrounding polygon plane are set as detection targets, and when these polygon planes have no intersection, all polygon planes are set as detection targets. An image processing method. A rearrangement step of rearranging the plurality of polygonal planes by brightness and hue angle;
3. The image processing method according to claim 1, wherein in the intersection detection step, detection is performed in order from a polygonal plane that approximates the brightness and hue angle of the color coordinates.   The mapping step searches for the nearest point from the polygon plane detected in the detection step and the polygon plane surrounding the polygon plane, and associates it with the nearest store. The image processing method according to item.   The image processing method according to any one of claims 1 to 4, wherein the predetermined point is a point on a gray axis of the second color space. The second device-dependent color is a device-dependent color of the printer;
A printer comprising a color conversion table created by the image processing method according to claim 1, and performing color conversion based on the color conversion table. A correspondence relationship between the first color space of the first device-dependent color and the second color space of the second device-dependent color is set, and the first color in the first color space is determined based on the correspondence relationship. An image processing device for converting to a second color in a second color space,
Color conversion table setting means for setting a color conversion table to be referred to when converting the first color into the second color based on a predetermined conversion rule;
Outer shell setting means for setting the outer shell of the second color space by combining a plurality of polygonal planes;
Detecting means for detecting a polygonal plane having an intersection of a half line from a predetermined point in the second color space to a color coordinate of a color converted based on the color conversion table;
An inside / outside determination means for comparing the intersection point with the color coordinate and determining whether the color coordinate is within the outer shell;
Mapping means for associating the color coordinates in the outer shell when the color coordinates are determined not to be in the outer shell by the inner / outer determination means;
Comprising
In the detecting means, first, the polygon plane detected immediately before and the surrounding polygon plane are set as detection targets, and when these polygon planes have no intersection, all polygon planes are set as detection targets. An image processing apparatus. The image processing apparatus further includes a graphic board equipped with a GPU (Graphics Processing Unit),
The detection means causes the GPU to detect the polygonal plane having an intersection with the half line,
The image processing apparatus according to claim 6, wherein the inside / outside determination unit causes the GPU to calculate a distance from the predetermined point to the color coordinate and a distance from the predetermined point to the intersection. A correspondence relationship between the first color space of the first device-dependent color and the second color space of the second device-dependent color is set, and the first color in the first color space is determined based on the correspondence relationship. An image processing program for causing a computer to realize a function of converting to a second color in a second color space,
A color conversion table setting function for setting a color conversion table to be referred to when converting the first color into the second color based on a predetermined conversion rule;
An outer shell setting function for setting an outer shell of the second color space by combining a plurality of polygonal planes;
A detection function for detecting a polygonal plane having an intersection of a half line from a predetermined point in the second color space to a color coordinate of a color converted based on the color conversion table;
An inside / outside determination function that compares the intersection point with the color coordinates and determines whether the color coordinates are within the outer shell;
A mapping function for associating the color coordinates in the outer shell when the color coordinates are determined not to be in the outer shell by the inner / outer determination function;
Is realized on a computer,
In the detection function, first, the polygon plane detected immediately before and the surrounding polygon planes are set as detection targets, and when these polygon planes have no intersection, all polygon planes are set as detection targets. An image processing program characterized by that.